1

.
QUANTUM FOAM, GRAVITY
AND
GRAVITATIONAL WAVES
Reginald T. Cahill
School of Chemistry, Physics and Earth Sciences
Flinders University
GPO Box 2100, Adelaide 5001, Australia
Reg.Cahill@inders.edu.au
Process Physics URL:
http://www.scieng.inders.edu.au/cpes/people/cahill r/processphysics.html
- September 2003 -
2
Abstract
It is shown that both the Newtonian and General Relativity theories for
gravity may be re-formulated as in-ow dynamics in which a substratum is
eectively absorbed by matter, with the gravitational force determined by
inhomogeneities of that ow. Analysis herein of the 1925-26 Dayton Miller
interferometer data reveals such a gravitational in-ow of space past the
Earth into the Sun. This data and that from the 1991 Roland DeWitte
coaxial cable experiment also suggests that the in-ow is turbulent, which
amounts to the observation of a gravitational wave phenomena. A gener-
alisation of the in-ow formalisms is proposed which passes all the tests
that General Relativity passed, but as well the new theory suggests that the
so-called spiral galaxy rotation-velocity anomaly may be explained without
the need of dark matter. As well analysis of data from the Michelson and
Morley, Miller, Illingworth, Jaseja et al, Torr and Kolen, and DeWitte ex-
periments reveal motion relative to the substratum. Special relativity eects
are caused by motion relative to the substratum. This implies that a new
ontology underlies the spacetime formalism.
PACS: 02.50.Ey, 04.60.-m,03.65.Bz
3
Contents
1 Introduction 3
2 In-Flow as Gravity 5
2.1 Newtonian Inow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Quantum Foam In-Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 Apparent Invariance of c . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 The Lorentz Transformation . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.5 The General Relativity Formalism . . . . . . . . . . . . . . . . . . . . . 13
2.6 General Relativity In-Flow . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.7 Generalised In-Flow - a New Theory of Gravity . . . . . . . . . . . . . . 16
2.8 The Dark Matter Eect . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.9 Gravity and Absolute Motion . . . . . . . . . . . . . . . . . . . . . . . . 18
2.10 Gravitational In-Flow and the GPS . . . . . . . . . . . . . . . . . . . . . 19
2.11 Gravitational Anomalies . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 Observations of Absolute Motion and In-Flow 21
3.1 Theory of the Michelson Interferometer . . . . . . . . . . . . . . . . . . . 21
3.2 The Michelson-Morley Experiment: 1887 . . . . . . . . . . . . . . . . . . . 25
3.3 The Miller Interferometer Experiment: 1925-1926 . . . . . . . . . . . . . . 29
3.4 In-ow from the Miller Data . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.5 The Illingworth Experiment: 1927 . . . . . . . . . . . . . . . . . . . . . . 33
3.6 The New Bedford Experiment: 1963 . . . . . . . . . . . . . . . . . . . . . 35
3.7 The DeWitte Experiment: 1991 . . . . . . . . . . . . . . . . . . . . . . . . 37
3.8 The Torr-Kolen Experiment: 1981 . . . . . . . . . . . . . . . . . . . . . . 41
3.9 Galactic In-ow and the CMB Frame . . . . . . . . . . . . . . . . . . . . . 42
3.10 Gravitational Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4 Conclusions 46
5 References 46
1 Introduction
The new information-theoretic Process Physics [1, 2, 3, 4, 5, 6] provides for the rst
time an explanation of space as a decohering quantum foam system in which gravity is
an inhomogeneous ow of the quantum foam into matter. As shown herein analysis of
data from various Michelson interferometer experiments has demonstrated that absolute
motion relative to space had been observed by Michelson and Morley [7], Miller [8],
Illingworth [9] and Jaseja et al [10] contrary to common belief within physics that absolute
motion has never been observed. The key discovery being that the presence of a gas is
required in order that a Michelson interferometer [11] be able to detect motion relative
to the quantum-foam substratum of space. This eect has gone unnoticed for over
4
100 years. All gas-mode Michelson interferometers have detected absolute motion but,
because the role of the gas had not been realised, the analysis of the data had been
incorrect, except for the experiment by Miller who cleverly developed a technique to
bypass the long-term deciency in understanding of the interferometer. Vacuum operated
Michelson interferometers are blind to absolute motion. This has also gone unnoticed
and has resulted in enormous confusion in the understanding of the experimental study of
relativistic eects. Here a comprehensive analysis of the above data is presented together
with the data from the non-interferometer experiments by Torr and Kolen [12], and by
DeWitte [13]. All these experiments agree on the direction and speed of absolute motion
of the solar system through the quantum-foam substratum.
The Dayton Miller extensive Michelson interferometer experimental data also reveals,
as shown here, the in-ow of space into the Sun which manifests as gravity, as well as the
orbital motion of the Earth about the Sun. The experimental data suggests that the in-
ow is turbulent, which amounts to the observation of a gravitational wave phenomena.
The DeWitte data also indicates a similar level of turbulence in the in-ow.
The extensive experimental data shows that absolute motion is consistent with rel-
ativistic eects. Indeed relativistic eects are caused by dynamical eects associated
with absolute motion, as proposed by Lorentz. The Lorentz transformation is seen to be
a consequence of absolute motion dynamics. Vacuum Michelson interferometer experi-
ments or its equivalent [14, 15, 16, 17, 18] cannot detect absolute motion, but their null
results do support this interpretation and form a part of the experimental predictions of
the new physics.
A new in-ow theory of gravity in the classical limit is proposed. It passes all the
standard tests that the Newtonian and the General Relativity theories of gravity have
passed, including the operation of the Global Positioning System. However it appears
that this new theory may explain as well the spiral galaxy rotation-velocity anomaly with-
out invoking dark matter. As well this new theory is expected to predict the turbulent
ow which is manifested in the existing experimental observations of absolute motion.
Other gravitational anomalies also now appear to be capable of being explained. These
developments amount to new physics.
This paper has two main sections, 2.In-Flow as Gravity which presents the origin
and properties of this new theory of gravity, and 3.Observations of Absolute Motion
and In-Flow which analyses the extensive data that supports this new theory of gravity.
Signicantly this new theory departs from both the Newtonian and General Relativity
theories in key aspects, and these experimental signatures are evident in the experimental
data. This new theory of gravity has stimulated new experiments to study in particular
the new gravitational wave phenomena. Because of the signicant development of our
understanding of how to detect absolute motion and ipso facto gravitational in-ows
these new experiments are basically bench-top experiments. One such experiment is
operating at Flinders university under the direction of Professor Warren Lawrance, and
a report of the analysis of the data will be soon forthcoming.
5
2 In-Flow as Gravity
2.1 Newtonian Inow
We begin here the analysis that will lead to the new theory and explanation of gravity.
In this theory gravitational eects are caused solely by an inhomogeneous ow of the
quantum foam. This is not a ow through space, but essentially a rearrangement of the
quantum-foam which globally is most easily described as a ow. This is a subtle aspect
of this new physics. The new information-theoretic concepts underlying this physics were
discussed in [1, 2]. Essentially matter eectively acts as a sink for that quantum foam.
To begin with it should be noted that even Newtonian gravity is suggestive of a ow
explanation of gravity. In that theory the gravitational acceleration g is determined by
the matter density according to
.g = 4G. (1)
For g = 0 this gravitational acceleration g may be written as the gradient of the
gravitational potential ,
g = , (2)
where the gravitational potential is now determined by
2
= 4G. Here, as usual, G
is the gravitational constant. Now as 0 we can choose to have 0 everywhere if
0 at innity. So we can introduce v
2
= 2 0 where v is some velocity vector
eld. Here the value of v
2
is specied, but not the direction of v. Then
g =
1
2
(v
2
) = (v.)v +v (v). (3)
For irrotational ow v = 0. Then g is the usual Euler expression for the acceleration
of a uid element in a time-independent or stationary uid ow. If the ow is time
dependent the Euler expression suggests the extra time-dependent term in
g = (v.)v +v (v) +
v
t
. (4)
This equation is then to be accompanied by the Newtonian equation for the ow eld
1
2

2
(v
2
) = 4G. (5)
While this hints at a uid ow interpretation of Newtonian gravity the fact that the
direction of v is not specied by (5) suggests that some generalisation is to be expected
in which the direction of v is specied. Of course within the uid ow interpretation
(4) and (5) are together equivalent to the Universal Inverse Square Law for Gravity.
Indeed for a spherically symmetric distribution of matter of total mass M the velocity
eld outside of the matter
v(r) =

2GM
r
r, (6)
6
satises (5) and reproduces the inverse square law form for g using (4):
g =
GM
r
2
r. (7)
The in-ow direction r in (6) may be replaced by any other direction, in which case
however the direction of g in (7) remains radial.
Of the many new eects predicted by the generalisation of (5), see section 2.7, one
is that this Inverse Square Law is only valid outside of spherically symmetric matter
systems. Then, for example, the Inverse Square Law is expected to be inapplicable to
spiral galaxies. The incorrect assumption of the universal validity of this law led to the
notion of dark matter in order to reconcile the faster observed rotation velocities of
matter within such galaxies compared to that predicted by the above law.
2.2 Quantum Foam In-Flow
To arrive at the new in-ow theory of gravity we require that the velocity eld v(r, t) be
specied and measurable with respect to a suitable frame of reference. We shall use the
Cosmic Microwave Background (CMB) frame of reference for that purpose [19]; see also
section 3.9. Then an object has velocity v
0
(t) = dr
0
(t)/dt with respect to that CMB
frame, where r
0
(t) is the position of the object wrt that frame. We then dene
v
R
(t) = v
0
(t) v(r
0
(t), t), (8)
as the velocity of the object relative to the quantum foam at the location of the object.
Process Physics leads to the Lorentzian interpretation of so called relativistic eects.
This means that the speed of light is only c wrt the quantum-foam system, and that time
dilation eects for clocks and length contraction eects for rods are caused by the motion
of clocks and rods relative to the quantum foam. So these eects are real dynamical
eects caused by the quantum foam. We conjecture that the path of an object through an
inhomogeneous and time-varying quantum-foam is determined by a variational principle,
namely the path r
0
(t) minimises the travel time (for early investigations of the in-ow
approach to gravity see Ives [20] and Kirkwood [21, 22]),
[r
0
] =
_
dt
_
1
v
2
R
c
2
_
1/2
, (9)
with v
R
given by (8). Under a deformation of the trajectory r
0
(t) r
0
(t) + r
0
(t),
v
0
(t) v
0
(t) +
dr
0
(t)
dt
, and we also have
v(r
0
(t) + r
0
(t), t) = v(r
0
(t), t) + (r
0
(t).)v(r
0
(t)) + ... (10)
Then
= [r
0
+ r
0
] [r
0
]
7
=
_
dt
1
c
2
v
R
.v
R
_
1
v
2
R
c
2
_
1/2
+ ...
=
_
dt
1
c
2
_
v
R
.(r
0
.)v v
R
.
d(r
0
)
dt
_
_
1
v
2
R
c
2
_
1/2
+ ...
=
_
dt
1
c
2
_
_
_
_
_
_
v
R
.(r
0
.)v

1
v
2
R
c
2
+ r
0
.
d
dt
v
R

1
v
2
R
c
2
_
_
_
_
_
_
+ ...
=
_
dt
1
c
2
r
0
.
_
_
_
_
_
_
(v
R
.)v +v
R
(v)

1
v
2
R
c
2
+
d
dt
v
R

1
v
2
R
c
2
_
_
_
_
_
_
+ ...
(11)
Hence a trajectory r
0
(t) determined by = 0 to O(r
0
(t)
2
) satises
d
dt
v
R

1
v
2
R
c
2
=
(v
R
.)v +v
R
(v)

1
v
2
R
c
2
. (12)
Let us now write this in a more explicit form. This will also allow the low speed limit to
be identied. Substituting v
R
(t) = v
0
(t) v(r
0
(t), t) and using
dv(r
0
(t), t)
dt
= (v
0
.)v +
v
t
, (13)
we obtain
d
dt
v
0

1
v
2
R
c
2
= v
d
dt
1

1
v
2
R
c
2
+
(v.)v v
R
(v) +
v
t

1
v
2
R
c
2
. (14)
Then in the low speed limit v
R
c we obtain
dv
0
dt
= (v.)v v
R
(v) +
v
t
= g(r
0
(t), t) + (v) v
0
, (15)
which agrees with the Newtonian form (4) for zero vorticity (v = 0). Hence (14) is a
generalisation of (4) to include Lorentzian dynamical eects, for in (14) we can multiply
both sides by the rest mass m
0
of the object, and then (14) involves
m(v
R
) =
m
0

1
v
2
R
c
2
, (16)
8
the so called relativistic mass, and (14) acquires the form
d
dt
(m(v
R
)v
0
) = F, (17)
where F is an eective force caused by the inhomogeneities and time-variation of the
ow. This is essentially Newtons 2nd Law of Motion in the case of gravity only. That
m
0
cancels is the equivalence principle, and which acquires a simple explanation in terms
of the ow. Note that the occurrence of 1/
_
1
v
2
R
c
2
will lead to the precession of the
perihelion of planetary orbits, and also to horizon eects wherever |v| = c: the region
where |v| < c is inaccessible from the region where |v| > c. Also (9), in conjunction with
(51), is easily used to show that the new theory of gravity agrees with that of General
Relativity for the operation of the GPS satellite navigation system, when the in-ow is
given by (6); see section 2.10.
Equation (9) involves various absolute quantities such as the absolute velocity of
an object relative to the quantum foam and the absolute speed c also relative to the
foam, and of course absolute velocities are excluded from the General Relativity (GR)
formalism. However (9) gives (with t x
0
0
)
d
2
= dt
2

1
c
2
(dr
0
(t) v(r
0
(t), t)dt)
2
= g

(x
0
)dx

0
dx

0
, (18)
which is the Panleve-Gullstrand [23, 24] form of the metric g

for GR. All of the above

is very suggestive that useful information for the ow dynamics may be obtained from
GR by restricting the choice of metric to the Panleve-Gullstrand form. We emphasize
that the absolute velocity v
R
has been measured, see section 3.4, and so this in-ow
theory of gravity is no longer speculative.
2.3 Apparent Invariance of c
The quantum foam induces actual dynamical time dilations and length contractions in
agreement with the Lorentz interpretation of special relativistic eects. As a consequence
of this observers in uniform motion through the foam will on measurement of the speed
of light obtain always the same numerical value c, so long as they do not adjust their
observational data to take account of these dynamical eects. So the special relativistic
eects are very much an aspect of physical reality, but nevertheless the absolute motion
causing these eects is observable.
To see this explicitly consider how various observers P, P

, .. moving with dierent

speeds through the foam, might measure the speed of light. They each acquire a standard
rod and an accompanying standardised clock. That means that these standard rods
would agree if they were brought together, and at rest with respect to the quantum
foam they would all have length l
0
, and similarly for the clocks. Observer P and
accompanying rod are both moving at speed v
R
relative to the quantum foam, with the
rod longitudinal to that motion, for simplicity. P then measures the time t
R
, with the
clock at end A of the rod, for a light pulse to travel from end A to the other end B and
back again to A. The light travels at speed c relative to the quantum-foam. Let the
9
time taken for the light pulse to travel from A B be t
AB
and from B A be t
BA
, as
measured by a clock at rest with respect to the quantum foam. The length of the rod
moving at speed v
R
is contracted to
l
R
= l
0

1
v
2
R
c
2
. (19)
In moving from A to B the light must travel an extra distance because the end B travels
a distance v
R
t
AB
in this time, thus the total distance that must be traversed is
ct
AB
= l
R
+ v
R
t
AB
, (20)
Similarly on returning from B to A the light must travel the distance
ct
BA
= l
R
v
R
t
BA
. (21)
Hence the total travel time t
0
is
t
0
= t
AB
+ t
BA
=
l
R
c v
R
+
l
R
c + v
R
(22)
=
2l
0
c

1
v
2
R
c
2
. (23)
Because of the time dilation eect for the moving clock
t
R
= t
0

1
v
2
R
c
2
. (24)
Then for the moving observer the speed of light is dened as the distance the observer
believes the light travelled (2l
0
) divided by the travel time according to the accom-
panying clock (t
R
), namely 2l
0
/t
R
= c. So the speed v
R
of the observer through
the quantum foam is not revealed by this procedure, and the observer is erroneously led
to the conclusion that the speed of light is always c. This invariance of c follows from
two or more observers in manifest relative motion all obtaining the same speed c by this
procedure. Despite this failure this special eect is actually the basis of the spacetime
measurement protocol. That this protocol is blind to the absolute motion has led to
enormous confusion within physics. However it is possible to overcome the blindness of
this procedure and to manifestly reveal an observers absolute velocity of motion v
R
. A
simple way to do this is shown in gure 1. This involves two identical antiparallel lasers.
Then the dierence in travel time through vacuum to the detector is
t =
L
c v
R

L
c + v
R
,
= 2
L
c
v
R
c
+ O(
v
2
R
c
2
). (25)
10
which is a 1st-order eect, and for that reason the time dilation and length contraction
eects have been neglected. Here for simplicity v
R
is along the axis of the instrument.
The speed v
R
is determined from the variation in beat frequency as the device is rotated.
The main technical diculty is in maintaining the frequency stability of the two lasers.
It is important to note that this device does not require synchronisation of the two clocks
(here lasers). If the two arms are placed at 90
0
to each other as in the New Bedford
experiment, see section 3.6, then the eect becomes null. To obtain a non-null eect in
this arrangement a gas is required in the air-paths. In the New Bedford experiment that
gas was in the masers. This is one of many experiments where the role of a gas in an
interferometer has played a critical but, until now, unrecognised role.
-
??
@@
Laser 1 Laser 2
L L
D
Figure 1: A 1st-order device for detecting absolute motion. Light from two identical
lasers is combined and their beat frequency is detected at D.
2.4 The Lorentz Transformation
Here we show that the real dynamical eects of absolute moton results in certain special
observational data being related by the Lorentz transformation. This involves the use
of the radar measurement protocol for acquiring observational space and time data of
distant events, and subsequently displaying that data in a spacetime construct. In this
protocol the observer records the time of emission and reception of radar pulses (t
r
> t
e
)
travelling through the space of quantum foam, and then retrospectively assigns the time
and distance of a distant event B according to (ignoring directional information for
simplicity)
T
B
=
1
2
(t
r
+ t
e
), D
B
=
c
2
(t
r
t
e
), (26)
where each observer is now using the same numerical value of c. The event B is then
plotted as a point in an individual geometrical construct by each observer, known as a
spacetime record, with coordinates (D
B
, T
B
). This is no dierent to a historian recording
events according to some agreed protocol. We now show that because of this protocol
and the quantum foam dynamical eects, observers will discover on comparing their
historical records of the same events that the expression

2
AB
= T
2
AB

1
c
2
D
2
AB
, (27)
is an invariant, where T
AB
= T
A
T
B
and D
AB
= D
A
D
B
are the dierences in times
and distances assigned to events A and B using the above measurement protocol (26),
11
so long as both are suciently small compared with the scale of inhomogeneities in the
velocity eld.
A
P(v
0
= 0)
B (t

B
)
D D
B
T
P

(v

0
)

*
H
H
H
H
H
H
H
H
H HY
t
e
T
B
t
r

Figure 2: Here T D is the spacetime construct (from the measurement protocol) of

a special observer P at rest wrt the quantum foam, so that v
0
= 0. Observer P

is
moving with speed v

0
as determined by observer P, and therefore with speed v

R
= v

0
wrt the quantum foam. Two light pulses are shown, each travelling at speed c wrt both
P and the quantum foam. As we see later these one-way speeds for light, relative to the
quantum foam, are equal by observation. Event A is when the observers pass, and is
also used to dene zero time for each for convenience.
To conrm the invariant nature of the construct in (27) one must pay careful attention
to observational times as distinct from protocol times and distances, and this must be
done separately for each observer. This can be tedious. We now demonstrate this for
the situation illustrated in gure2.
By denition the speed of P

according to P is v

0
= D
B
/T
B
and so v

R
= v

0
, where
T
B
and D
B
are the protocol time and distance for event B for observer P according to
(26). Then using (27) P would nd that (
P
AB
)
2
= T
2
B

1
c
2
D
2
B
since both T
A
= 0 and
D
A
=0, and whence (
P
AB
)
2
= (1
v
2
R
c
2
)T
2
B
= (t

B
)
2
where the last equality follows from
the time dilation eect on the P

clock, since t

B
is the time of event B according to that
clock. Then T
B
is also the time that P

would compute for event B when correcting for

the time-dilation eect, as the speed v

R
of P

through the quantum foam is observable

by P

. Then T
B
is the common time for event B assigned by both observers. For P

we obtain directly, also from (26) and (27), that (

P

AB
)
2
= (T

B
)
2

1
c
2
(D

B
)
2
= (t

B
)
2
, as
D

B
= 0 and T

B
= t

B
. Whence for this situation
(
P
AB
)
2
= (
P

AB
)
2
, (28)
12
and so the construction (27) is an invariant.
While so far we have only established the invariance of the construct (27) when one
of the observers is at rest wrt to the quantum foam, it follows that for two observers
P

and P

both in motion wrt the quantum foam it follows that they also agree on the
invariance of (27). This is easily seen by using the intermediate step of a stationary
observer P:
(
P

AB
)
2
= (
P
AB
)
2
= (
P

AB
)
2
. (29)
Hence the protocol and Lorentzian eects result in the construction in (27) being indeed
an invariant in general. This is a remarkable and subtle result. For Einstein this in-
variance was a fundamental assumption, but here it is a derived result, but one which
is nevertheless deeply misleading. Explicitly indicating small quantities by prexes,
and on comparing records retrospectively, an ensemble of nearby observers agree on the
invariant

2
= T
2

1
c
2
D
2
, (30)
for any two nearby events. This implies that their individual patches of spacetime records
may be mapped one into the other merely by a change of coordinates, and that collectively
the spacetime patches of all may be represented by one pseudo-Riemannian manifold,
where the choice of coordinates for this manifold is arbitrary, and we nally arrive at the
invariant

2
= g

(x)x

, (31)
with x

= {T, D
1
, D
2
, D
3
}. For at metrics (31) is invariant under the well known
Lorentz transformation,
x

= L(v)

, (32)
where, for motion only in the x-direction,
x = (x

ct

)
ct = (ct

)
y = y

z = z

(33)
where = v/c and = 1/
_
1
2
. Here, in general, v is the relative velocity of the two
observers, determined by using the measurement protocol. The special feature of this
mapping between the observers spacetime records is that it does not involve the absolute
velocity of either observer relative to the quantum-foam substratum - their absolute
velocities. This feature was responsible for the rst two assumptions in (34). This
feature has caused enormous confusion in physics. It erroneously suggests that absolute
motion is incompatible with relativistic eects - that the observation of absolute motion
must be in conict with the observation of relativistic eects. For that reason reports
of the ongoing detection of absolute motion has been banned in physics for nearly 100
years. However to the contrary absolute motion and special relativistic eects are both
needed to understand and analyse the extensive experimental data reported in section 3.
The key insight is that absolute motion dynamically causes the time dilation and length
13
contraction eects. Without absolute motion there would be no special relativistic eects.
This insight runs counter to nearly 100 years of conventional wisdom within physics.
2.5 The General Relativity Formalism
The general relativity formalism is well known. It was constructed by Hilbert and Ein-
stein by amalgamating the special relativity invariance and, in the low speed limit, the
Newtonian theory of gravity. This resulted in the need for the key feature of employing
a non-at spacetime manifold. The three key assumptions were:
(1) The laws of physics have the same form in all
inertial reference frames.
(2) Light propagates through empty space with a
speed c independent of the speed of the
(a) source or (b) observer.
(3) In the limit of low speeds the new formalism
should agree with Newtonian gravity. (34)
The rst two assumptions, apart from 2(a) which remains completely valid, have
restricted truth in that they refer to the dynamical eects of absolute motion, and how
those eects enter into the description of physical phenomena when not correcting for the
eects of the absolute motion on the observers measuring clocks and rods. As we shall see
the third assumption is actually the weakest for we shall see that the Newtonian theory
of gravity was formulated under very special conditions; namely ones of high spherical
symmetry. When that symmetry is not present then Newtonian gravity is awed. There
is abundant experimental evidence to support this claim. Hence the weakest part of the
general relativity formalism is actually its link to Newtonian gravity. Nevertheless there
is something that is partially correct within the formalism for it has passed a number
of key tests, albeit with most tests occuring also in cases of high spherical symmetry, as
explained later. And so the aw in general relativity like that of the Newtonian theory
has essentially gone unnoticed. Here we analyse the general relativity formalism in order
to discover which aspect of it is actually responsible for its few successes. We shall see
that in fact in those cases it may be reformulated as an in-ow formalism.
From the above assumptions the equations which specify the metric tensor g

(x) of
the spacetime construct may be found to be
G

1
2
Rg

=
8G
c
2
T

, (35)
where G

is known as the Einstein tensor, T

is the energy-momentum tensor, R

=
R

and R = g

and g

is the matrix inverse of g

. The curvature tensor is

R

,
+

, (36)
14
where

is the ane connection

=
1
2
g

_
g

+
g

_
. (37)
In this formalism the trajectories of test objects are determined by

dx

d
dx

d
+
d
2
x

d
2
= 0, (38)
which is equivalent to minimising the functional
[x] =
_
dt

dx

dt
dx

dt
, (39)
wrt to the path x[t].
For the case of a spherically symmetric mass a solution of (35) for g

outside of that
mass M is the Schwarzschild metric
d
2
= (1
2GM
c
2
r
)dt
2

1
c
2
r
2
(d
2
+ sin
2
()d
2
)
dr
2
c
2
(1
2GM
c
2
r
)
. (40)
This solution is the basis of various experimental checks of General Relativity in which
the spherically symmetric mass is either the Sun or the Earth. The four tests are: the
gravitational redshift, the bending of light, the precession of the perihelion of Mercury,
and the time delay of radar signals. To these we should add the operation of the GPS;
see section 2.10.
However the solution (40) is in fact completely equivalent to the in-ow interpretation
of Newtonian gravity. Making the change of variables t t

and r r

= r with
t

= t +
2
c

2GMr
c
2

4GM
c
2
tanh
1

2GM
c
2
r
, (41)
the Schwarzschild solution (40) takes the form
d
2
= dt
2

1
c
2
(dr

2GM
r

dt

)
2

1
c
2
r
2
(d
2
+ sin
2
(

)d

), (42)
which is exactly the Panleve-Gullstrand form of the metric g

[23, 24] in (18) with the

velocity eld given exactly by the Newtonian form in (6). In which case the trajectory
equation (38) of test objects in the Schwarzschild metric is equivalent to solving (14).
Thus the minimisation of the functional in (39) is equivalent to the minimisation of
the functional in (9). This choice of coordinates corresponds to a particular frame of
reference in which the test object has velocity v
R
= vv
0
relative to the in-ow eld v.
It is conventional wisdom for practitioners in General Relativity to regard the choice
of coordinates or frame of reference to be entirely arbitrary and having no physical
signicance: no observations should be possible that can detect and measure v
R
. This
15
wisdom is based on two beliefs (i) that all attempts to detect v
R
, namely the detection of
absolute motion, have failed, and that (ii) the existence of absolute motion is incompatible
with the many successes of both the Special Theory of Relativity. Both of these beliefs
are demonstrably false.
The results in this section suggest, just as for Newtonian gravity, that General Rel-
ativity is nothing more than the dynamical equations for a velocity ow eld v(r, t),
atleast in those cases where it has been checked.
2.6 General Relativity In-Flow
Here we extract from General Relativity the in-ow formalism. To do this we must
clearly adopt the Panleve-Gullstrand form of the metric g

as that corresponding to the

observable quantum foam system, namely to an observationally detected special frame
of reference. This form for the metric involves a general velocity eld v(r, t) where for
precision we consider the coordinates r, t as that of observers at rest with respect to the
CMB frame. Note that in this frame v(r, t) is not necessarily zero, for mass acts as a
sink for the ow. We therefore merely substitute the metric
d
2
= g

dx

dx

= dt
2

1
c
2
(dr(t) v(r(t), t)dt)
2
, (43)
into (35) using (37) and (36). This metric involves the arbitrary time-dependent velocity
eld v(r, t). This is a very tedious computation and the results below were obtained by
using the symbolic mathematics capabilities of Mathematica. The various components
of the Einstein tensor are then
G
00
=

i,j=1,2,3
v
i
G
ij
v
j
c
2

j=1,2,3
G
0j
v
j
c
2

i=1,2,3
v
i
G
i0
+ c
2
G
00
,
G
i0
=

j=1,2,3
G
ij
v
j
+ c
2
G
i0
, i = 1, 2, 3.
G
ij
= G
ij
, i, j = 1, 2, 3. (44)
where the G

are given by
G
00
=
1
2
((trD)
2
tr(D
2
)),
G
i0
= G
0i
=
1
2
((v))
i
, i = 1, 2, 3.
G
ij
=
d
dt
(D
ij

ij
trD) + (D
ij

1
2

ij
trD)trD

1
2

ij
tr(D
2
) (D D)
ij
, i, j = 1, 2, 3. (45)
Here
D
ij
=
1
2
(
v
i
x
j
+
v
j
x
i
) (46)
16
is the symmetric part of the rate of strain tensor
v
i
x
j
, while the antisymmetric part is

ij
=
1
2
(
v
i
x
j

v
j
x
i
). (47)
In vacuum, with T

= 0, we nd from (35) and (44) that G

= 0 implies that G

= 0.
It is then easy to check that the in-ow velocity eld (6) satises these equations. This
simply expresses the previous observation that this Newtonian in-ow is completely
equivalent to the Schwarzschild metric. That the Scwarzschild metric in (40) is nothing
more than the Newtonian inverse square law (7) in disguise appears to be poorly known.
We note that the vacuum equations G

= 0 do not involve the speed of light; it appears

only in (44). It is therefore suggested that (44) amounts to the separation of the mea-
surement protocol, which involves c, from the supposed dynamics of gravity within the
General Relativity formalism, and which does not involve c. However the details of the
vacuum dynamics in (45) have not actually been tested: All the key tests of General Rel-
ativity are now seen to amount to a test only of [x]/x

= 0, which is the minimisation

of (9), when the in-ow eld is given by (44), and which is nothing more than Newtonian
gravity. Of course Newtonian gravity was itself merely based upon observations within
the Solar system, and this may have been too special to have revealed key aspects of
gravity. Hence, despite popular opinion, the General Relativity formalism is apparently
based upon rather poor evidence.
2.7 Generalised In-Flow - a New Theory of Gravity
Despite the limited insight into gravity which General Relativity is now seen to amount
to, here we look for possible generalisations of Newtonian gravity and its in-ow inter-
pretation by examining some of the mathematical structures that have arisen in (45).
For the case of zero vorticity v = 0 we have
ij
= 0 and also that we may write
v = u where u(r, t) is a scalar eld, and only one equation is required to determine u.
To that end we consider the trace of G
ij
. Note that tr(D) = .v, and that
d(.v)
dt
= (v.)(.v) +
(.v)
t
. (48)
Then using the identity
(v.)(.v) =
1
2

2
(v
2
) tr(D
2
)
1
2
(v)
2
+v.(v), (49)
and imposing

i=1,2,3
G
ii
= 8G, (50)
we obtain

t
(.v) +
1
2

2
(v
2
) +
1
4
((trD)
2
tr(D
2
)) = 4G. (51)
This is seen to be a possible generalisation of the Newtonian equation (5). Note that
General Relativity has suggested exactly the time derivative of the form suggested by
17
the Euler uid ow acceleration in (4) (see also (52)), and also the new term C(v) =
1
4
((trD)
2
tr(D
2
)). First note that for the case of the Solar system, with the mass
concentrated in one object, namely the Sun, we see that the in-ow eld (6) satises (51)
since in this special case C(v) = 0. As we shall see later the presence of the C term is also
well hidden when we consider the Earths gravitational eects, although there are various
known anomalies that indicate that a generalisation of Newtonian gravity is required.
Hence (51) in the case of the Solar system is indistinguishable from Newtonian gravity,
or the Schwarzschild metric within the General Relativity formalism, so long as we use
(9), in being able to determine trajectories of test objects. Hence (51) is automatically
in agreement with most of the so-called checks on Newtonian gravity and later General
Relativity. Note that (51) does not involve the speed of light c. Nevertheless we have not
derived (51)) from the underlying Quantum Homotopic Field Theory which arises from
the information-theoretic theory in [1], and indeed it is not a consequence of General
Relativity, as the G
00
equation of (45) requires that C(v) = 0 in vacuum. Equation (51)
at this stage should be regarded as a conjecture which will permit the exploration of
possible quantum-foam physics, at the classical level, and also allow comparison with
experiment.
As well we should comment on two other tests of General Relativity. One is the
observed decay of the orbits of binary pulsars. From (16) with the in-ow (6) it is easily
seen that circular orbits are stable. However for elliptical orbits not only is there a
precession of the orbit but the orbit is not stable. On dimensional grounds we would
expect a decay rate of the magnitude observed for binary pulsars. The other test is
the prediction of the cosmological curvature of the universe and associated with the Big
Bang. As noted in [1] process physics also predicts a growing non-at universe. These
cosmological aspects are clearly not included in (51), which is only applicable to local
eects.
However one key aspect of (51) should be noted here, namely that being a non-linear
uid-ow dynamical system we would expect the ow to be turbulent, particularly when
the matter is not spherically symmetric or inside even a spherically symmetric distribu-
tion of matter, since then the C(v) term is non-zero and it will drive that turbulence.
In the following sections we shall see that the experiments that reveal absolute motion
also reveal evidence of turbulence.
2.8 The Dark Matter Eect
Because of the C(v) term (51) would predict that the Newtonian inverse square law would
not be applicable to systems such as spiral galaxies, because of their highly non-spherical
distribution of matter. Of course attempts to retain this law, despite its manifest failure,
has led to the spurious introduction of the notion of dark matter within spiral galaxies,
and also at larger scales. From
g =
1
2
(v
2
) +
v
t
, (52)
18
which is (4) for irrotational ow, we see that (51) gives
.g = 4G C(v), (53)
and taking running time averages to account for turbulence
.<g>= 4G <C(v)>, (54)
and writing the extra term as <C(v)>= 4G
DM
we see that
DM
would act as an
eective matter density, and it is suggested that it is the consequences of this term which
have been misinterpreted as dark matter. Here we see that this eect is actually the
consequence of quantum foam eects within the new proposed dynamics for gravity, and
which becomes apparent particularly in spiral galaxies. Note that (51) is an equation for
v, and now involves the direction of v, unlike the special case of Newtonian gravity (5).
Because v = 0 we can write (51) in the form
v(r, t) =
1
4
_
t
dt

_
d
3
r

(r r

)
1
2

2
(v
2
(r

, t

)) + 4G(r

, t

) + C(v(r

, t

))
|r r

|
3
, (55)
which allows the determination of the time evolution of v.
The new ow dynamics encompassed in (51) thus accounts for most of the known
gravitational phenomena, but will lead to some very clear cut experiments that will dis-
tinguish it from the two previous attempts to model gravitation. It turns out that these
two attempts were based on some key accidents of history. In the case of the Newtonian
modelling of gravity the prime accident was of course the Solar system with its high de-
gree of spherical symmetry. In each case we had test objects, namely the planets, in orbit
about the Sun, or we had test object in orbit about the Earth. In the case of the General
Relativity modelling the prime accident was the mis-reporting of the Michelson-Morley
experiment, and the ongoing belief that the so called relativistic eects are incompati-
ble with absolute motion. We shall consider in detail later some further anomalies that
might be appropriately explained by this new modelling of gravity. Of course that the
in-ow has been present in various experimental data is also a signicant argument for
something like (51) to model gravity. Key new experimental techniques will enable the
consequences of (51) to be tested. If necessary these experiments will provide insights
into possible modications to (51).
2.9 Gravity and Absolute Motion
We consider here why the existence of absolute motion and as well the presence of the
C(v) term appears to have escaped attention in the case of gravitational experiments
and observations near the Earth, despite the fact, in the case of the C(v) term, that the
presence of the Earth breaks the spherical symmetry of the matter distribution of the
Sun.
First note that if we have a matter distribution (r) at rest in the space of quantum
foam, and that (51) has solution v
0
(r, t), and then with g
0
(r, t) given by (52), then
19
when the same matter distribution is uniformly translating at velocity V, that is (r)
(r Vt), then a solution to (51) is
v(r, t) = v
0
(r Vt, t) +V. (56)
Note that this is a manifestly time-dependent process and the time derivative in (4) or
(14) and (51) plays an essential role. As well the result is nontrivial as (51) is a non-linear
equation. The solution (56) follows because (i) the expression for the acceleration g(r, t)
gives, and this expression occurs in (51),
g(r, t) =
v
0
(r Vt, t)
t
+ ((v
0
(r Vt, t) +V).)(v
0
(r Vt, t) +V),
=
v
0
(r Vt

, t)
t

t
+g
0
(r Vt, t) + (V.)v
0
(r Vt, t),
= (V.)v
0
(r Vt, t) +g
0
(r Vt, t) + (V.)v
0
(r Vt, t),
= g
0
(r Vt, t), (57)
as there is a key cancellation of two terms in (57), and (ii) clearly C(v
0
(rVt, t) +V) =
C(v
0
(r Vt, t)), and so this term is also simply translated. Hence apart from the
translation eect the acceleration is the same. Hence the velocity vector addition rule in
(56) is valid for generating the vector ow eld for the translating matter distribution.
This is why the large absolute motion velocities of some 400 km/s do not interfer with
the usual computation and observation of gravitational forces.
For Earth based gravitational phenomena the motion of the Earth takes place within
the velocity in-ow towards the Sun, and the velocity sum rule (56) is only approximately
valid as now V V(r, t) and no longer corresponds to uniform translation, and manifests
turbulence. To be a valid approximation the inhomogeneity of V(r, t) must be much
smaller than that of v
0
(r Vt, t), which it is, as the Earths centripetal acceleration
about the Sun is approximately 1/1000 that of the Earths gravitational acceleration
at the surface of the Earth. Nevertheless turbulence associated with the C(v) term is
apparent in experimental data. The validity of this approximation demonstrates that
the detection of a cosmic absolute motion and the in-ow theory of gravity are consistent
with the older methods of computing gravitational forces. This is why both the presence
of the C(v) term, the in-ow and the absolute motion have gone almost unnoticed in
Earth based gravitational experiments, except for various anomalies; see section 2.11.
2.10 Gravitational In-Flow and the GPS
We show here that the new in-ow theory of gravity together with the observed absolute
velocity of motion of the solar system through space are together compatible with the
operation of the Global Positioning System (GPS). This turns out to be an almost trivial
exercise. As usual in this system the eects of the Sun and Moon are neglected. Various
eects need to be included as the system relies upon extremely accurate atomic clocks
in the satellites forming the GPS constellation. Within both the new theory and general
relativity these clocks are eected by both their speed and the gravitational eects of
20
the Earth. As well the orbits of these satellites and the trajectories of radio signals from
the satellites need to be computed. For the moment we assume spherical symmetry for
the Earth. The eects of non-sphericity will be discussed below. In general relativity
the orbits and signalling time delays are determined by the use of the geodesic equation
(38) and the Schwarzschild metric (40). However these two equations are equivalent
to the orbital equation (16) and the velocity eld (56), with a velocity V of absolute
motion, and with the in-ow given by (6), noting the result in section 2.9. For EM
signalling the elapsed time in (9) requires careful treatment. Hence the two systems are
completely mathematically equivalent: the computations within the new system may
most easily be considered by relating them to the mathematically equivalent general
relativity formalism. There are nevertheless two possible dierences between the two
theories. One is their dierent treatment of the non-sphericity of the Earth particularly
via the C(v) term, and (2) the eects of the in-ow turbulence. It is possible that these
eects could lead to new experimental comparisons of the two theories, as well as perhaps
to an improved accuracy within the system if these new eects are large enough.
2.11 Gravitational Anomalies
As noted in section 2.1 Newtons Inverse Square Law of Gravitation may only be strictly
valid in cases of spherical symmetry. The theory that gravitational eects arise from
inhomogeneities in the quantum foam ow implies that there is no universal law of
gravitation because the inhomogeneities are determined by non-linear uid equations
and the solutions have no form which could be described by a universal law. Funda-
mentally there is no generic uid ow behaviour. The Inverse Square Law is then only an
approximation, with large deviations expected in the case of spiral galaxies. Nevertheless
Newtons gravitational constant G will have a denite value as it quanties the eective
rate at which matter dissipates the information content of space.
From these considerations it follows that the measurement of the value of G will be
dicult as the measurement of the forces between two of more objects, which is the usual
method of measuring G, will depend on the geometry of the spatial positioning of these
objects in a way not previously accounted for because the Newtonian Inverse Square Law
has always been assumed, or in some case a specied change in the form of the law has
been used. But in all cases a law has been assumed, and this may have been the aw
in the analysis of data from such experiments. This implies that the value of G from
such experiments will show some variability as a systematic eect has been neglected in
analysing the experimental data, for in none of these experiments is spherical symmetry
present. So experimental measurements of G should show an unexpected contextuality.
As well the inuence of surrounding matter has also not been properly accounted for. Of
course any eects of turbulence in the inhomogeneities of the ow has presumably also
never even been contemplated.
The rst measurement of G was in 1798 by Cavendish using a torsional balance. As
the precision of experiments increased over the years and a variety of techniques used the
disparity between the values of G has actually increased. In 1998 CODATA increased
the uncertainty in G from 0.013% to 0.15%. One indication of the contextuality is that
21
measurements of G produce values that dier by nearly 40 times their individual error
estimates [26]. It is predicted that these G anomalies will only be resolved when the new
theory of gravity is used in analysing the data from these experiments, and that these
precision G experiments provide another opportunity to check the new theory of gravity.
3 Observations of Absolute Motion and In-Flow
Absolute motion is motion relative to space itself. Absolute motion suggests that space
has some structure, and indeed evidence of such structure has been repeatedly discov-
ered over the last 115 years. It turns out that Michelson and Morley in their historic
experiment of 1887 did detect absolute motion, but rejected their own ndings because
using Galilean relativity the determined speed of some 8 km/s was less than the 30
km/s orbital speed of the Earth. The data was clearly indicating that the theory for
the operation of the Michelson interferometer was not adequate. Rather than reaching
this conclusion Michelson and Morley came to the incorrect conclusion that their results
amounted to the failure to detect absolute motion. This had an enormous impact on
the development of physics, for as is well known Einstein accepted the erroneous evi-
dence for the absence of absolute motion eects in his reinterpretion of the then extant
Lorentzian interpretation. By the time Miller had nally gured out how to use and
properly analyse data from his Michelson interferometer absolute motion had become a
forbidden concept within physics, as it still is at present. The experimental observations
by Miller and others of absolute motion has continued to be scorned and rejected by the
physics community. Fortunately as well as revealing absolute motion the experimental
data also reveals evidence in support of a new theory of gravity.
3.1 Theory of the Michelson Interferometer
We now show for the rst time in over 100 years how three key eects together permit
the Michelson interferometer [11] to reveal the phenomenon of absolute motion when
operating in the presence of a gas, with the third eect only discovered in 2002 [5].
The main outcome is the expression for the time dierence for light travelling via the
orthogonal arms,
t = k
2
L|v
P
|
2
c
3
cos(2( )) + O(|v
P
|
4
). (58)
Here v
P
is the projection of the absolute velocity v of the interferometer through the
quantum-foam onto the plane of the interferometer, where the projected velocity vector
v
P
has azimuth angle relative to the local meridian, and is the angle of one arm from
that meridian, i.e. the arm has angle to the projected direction of motion. The k
2
factor is k
2
= (n
2
1) where n is the refractive index of the gas through which the light
passes, and where we have assumed that n 1
+
, L is the rest-frame length of each arm
and c is the speed of light relative to the quantum foam in the absence of a gas. This
expression requires considerable care in its derivation, and here only a simplied analysis
will be given for the case when the arms are either parallel or orthogonal to the direction
22
- -

6
?
?

L
A B L
C
D
- - -
-

C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
CW

A
1
A
2
D
B
C
v
(a) (b)
Figure 3: Schematic diagrams of the Michelson Interferometer, with beamsplitter/mirror
at A and mirrors at B and C on arms from A, with the arms of equal length L when at
rest. D is a quantum detector that causes localisation of the photon state by a collapse
process. In (a) the interferometer is at rest in space. In (b) the interferometer is moving
with speed v relative to space in the direction indicated. Interference fringes are observed
at the quantum detector D. If the interferometer is rotated in the plane through 90
o
, the
roles of arms AC and AB are interchanged, and during the rotation shifts of the fringes
are seen in the case of absolute motion, but only if the apparatus operates in a gas. By
counting fringe changes the speed v may be determined.
of motion and when the direction of motion is in the plane of the interferometer. The
expression in (58) actually follows from three key eects: (i) the dierence in geometrical
length of the two paths when the interferometer is in absolute motion, as rst realised
by Michelson, (ii) the Fitzgerald-Lorentz contraction of the arms along the direction of
motion, and (iii) that these two eects precisely cancel in vacuum, but leave a residual
eect if operated in a gas, because the speed of light through the gas is reduced compared
to vacuum. As well we shall take account of a fourth eect, namely the Fresnel drag in
the gas caused by its absolute motion.
The time dierence t is revealed by the fringe shifts on rotating the interferometer.
However another eect needs to be considered. This time dierence arises for light
generated by atomic transitions in a light source that is travelling with the interferometer.
And so there is a time dilation eect for this source. It turns out that fortunately because
of the high speed and the direction of the observed absolute motion, compared to the
orbital and in-ow velocities, that this eect is negligible as the change in the total v
2
over a year is suciently small. What is detected is the change in the projection of the
total velocity onto the plane of the interferometer both during a day, and also seasonally
due to the inclination of the plane of the ecliptic - the orbital plane, to the plane of
motion of the interferometer due to the Earths daily rotation on its axis. However it
should be noted that in the Kennedy-Thorndike [15] the eect of the absolute motion on
the frequency of the light source was overlooked. This resulted in an erroneous analysis
of data that was entirely instrumental noise.
In Newtonian physics, that is with no Fitzgerald-Lorentz contraction, k
2
= n
3
1
for gases, while in Einsteinian physics k = 0 reecting the fundamental assumption that
absolute motion is not measurable and indeed has no meaning. For air n = 1.00029,
23
and so k = 0.0241 and k
2
= 0.00058, which is close to the Einsteinian value of k = 0,
particularly in comparison to the Newtonian value of k = 1.0. This small but non-zero
k value explains why the Michelson interferometer experiments gave such small fringe
shifts. Fortunately it is possible to check the n dependence of k as one experiment [9]
was done in Helium gas, and this has an n
2
1 value signicantly dierent from that of
air.
As shown in gure 3 the beamsplitter/mirror when at A sends a photon (t) into a
superposition (t) =
1
(t) +
2
(t), with each component travelling in dierent arms of
the interferometer, until they are recombined in the quantum detector which results in
a localisation process, and one spot in the detector is produced. Repeating with many
photons reveals that the interference between
1
and
2
at the detector results in fringes.
These fringes actually only appear if the mirrors are not quite orthogonal, otherwise the
screen has a uniform intensity and this intensity changes as the interferometer is rotated,
as shown in the analysis by Hicks [25]. To simplify the analysis here assume that the two
arms are constructed to have the same lengths L when they are physically parallel to
each other and perpendicular to v. Consider the Michelson interferometer operating in
a gas which is moving with the interferometer at speed v. The motion of the gas relative
to space results in a Fresnel drag eect. For simplicity consider only the cases when the
arms are parallel/orthogonal to the direction of motion, as shown in gure 3. Let the
arms have equal lengths L when at rest. The Fitzgerald-Lorentz relativistic eect is that
the arm AB parallel to the direction of motion is shortened to
L

= L

1
v
2
c
2
(59)
by absolute motion, while the length L of the transverse arm is unaected. We work
in the absolute rest frame. Consider the photon states in the AB arm. They travel at
speed V = c/nbv relative to the quantum-foam which is space, where n is the refractive
index of the gas and c is the speed of light in vacuum and relative to the space. Here
b = 1 1/n
2
is the Fresnel drag coecient which is well established experimentally. The
motion of the gas through the quantum foam slightly drags the light. The eect on the
speed is bv depending on the direction of the light relative to the direction of absolute
motion. Then the total travel time t
ABA
is
t
ABA
= t
AB
+ t
BA
=
L

c
n
+ bv v
+
L

c
n
bv + v
(60)
=
2Ln
c

1
v
2
c
2
1
1
v
2
n
2
c
2
. (61)
For the orthogonal arm we have by Pythagoras theorem
(V t
AC
)
2
= L
2
+ (vt
AC
)
2
. (62)
The speed V of light travelling from A to C (and also from C to A) is
V =
c
n
+ bv cos(), (63)
24
where is the angle of the transverse light path to the direction of motion of the inter-
ferometer, as shown in gure 3, and is given by
cos() =

1
L
2
(V t)
2
. (64)
Solving (62), (63) and (64) for V we obtain
V =
1
2
_
_
c
2
n
2
+

c
2
n
2
+ 4bv
2
_
_
. (65)
Then (62) gives t
AC
, and we obtain, with t
ACA
= t
AC
+ t
CA
= 2t
AC
, and for v c
t
0
0
90
0 = 2(t
ABA
t
ACA
) = 2
(n
2
1)(2 n
2
)L
nc
v
2
c
2
+ O(v
4
), (66)
for the change in relative travel times when the apparatus is rotated through 90
0
. The
factor of 2 arises because then the role of each arm is interchanged. For gases n 1
+
and we obtain
t
0
0
90
0 2
(n
2
1)L
c
v
2
c
2
+ O(v
4
). (67)
A more general analysis shows that when the arm AB has angle relative to the
projection of the velocity of absolute motion we obtain (58). Then on rotation through
90
0
the factor cos(2( )) changes by 2, so giving (58) the factor of 2 seen in (67).
The major signicance of this result is that this time dierence is not zero when a gas is
present in the interferometer, as conrmed by all gas-mode interferometer experiments.
Of course this result also shows that vacuum-mode experiments, with n = 1, will give
null results, as also conrmed by experiment [14, 15, 16, 17, 18]. So gas-mode Michelson
interferometers are blind to the eects of absolute motion, but they play a key role in
conrming the Fitzgerald-Lorentz contraction eect, and by using vacuum they separate
this eect from the refractive index eect.
It was Miller who rst introduced the parameter k as he appreciated that the op-
eration of the Michelson interferometer was not fully understood, although of course he
never realised that k is related to the refractive index of the gas present in the inter-
ferometer. This is very fortunate since being a multiplicative parameter a re-scaling of
old analyses is all that is required. t is non-zero when n = 1 because the refractive
index eect results in incomplete cancellation between the geometrical eect and the
Fitzgerald-Lorentz contraction eect. This incomplete cancellation arises whether we
include the Fresnel drag eect or not, so its role in gas-mode Michelson interferometers
is not critical. Leaving it out simply changes the overall sign in (58). Of course it was
this cancellation eect that Fitzgerald and Lorentz actually used to arrive at the length
contraction hypothesis, but they failed to take the next step and note that the can-
cellation would be incomplete in a gas operated Michelson interferometer. In a bizarre
development modern Michelson interferometer experiments use resonant vacuum cavities
rather than interference eects, but for which the analysis here is easily adapted, and
25
with the same consequences. That denies these experiments the opportunity to see abso-
lute motion eects. Nevertheless the experimentalists continue to misinterpret their null
results as evidence against absolute motion. Of course these experiments are therefore
restricted to merely checking the Fitzgerald-Lorentz contraction eect, and this is itself
of some interest.
All data from gas-mode interferometer experiments, except for that of Miller, has
been incorrectly analysed using only the rst eect as in Michelsons initial theoretical
treatment, and so the consequences of the other two eects have been absent. Repeating
the above analysis without these two eects we arrive at the Newtonian-physics time
dierence which, for v << V and n 1
+
, is
t = L
|v
P
|
2
c
3
cos(2( )) + O(|v
P
|
4
), (68)
that is k = 1. The value of t, which is typically of order 10
17
10
16
s in gas-mode
interferometers corresponding to a fractional fringe shift, is deduced from analysing the
fringe shifts, and then the speed v
M
has been extracted using (68), instead of the correct
form (67) (by the M subscript we indicate that the speed was obtained by using the
incorrect Michelson theory (68)). However it is very easy to correct for this oversight.
From (67) and (68) we obtain for the corrected absolute (projected) speed v
P
through
space, and for n 1
+
,
v
P
=
v
M

n
2
1
. (69)
For air the correction factor in (69) is signicant, and even more so for Helium.
3.2 The Michelson-Morley Experiment: 1887
Michelson and Morley reported that their interferometer experiment in 1887 gave a null-
result which since then, with rare exceptions, has been claimed to support the assump-
tion that absolute motion has no meaning. However to the contrary the Michelson-Morley
published data [7] shows non-null eects, but much smaller than they expected. They
made observations of thirty-six 360
0
turns using an L = 11 meter length interferometer,
achieved using multiple reections, operating in air in Cleveland (Latitude 41
0
30

N)
with six turns near 12 :00 hrs (7:00 hrs ST) on each day of July 8, 9 and 11, 1887 and
similarly near 18 :00 hrs (13:00 hrs ST) on July 8, 9 and 12, 1887. Each turn took ap-
proximately 6 minutes as the interferometer slowly rotated oating on a tank of mercury.
They published and analysed the average of each of the 6 data sets. The fringe shifts
were extremely small but within their observational capabilities.
The orientation of the stone slab base is indicated by marks 16, 1, 2, .., as in gure 4;
North is mark 16. The dominant eect was a uniform fringe drift caused by temporal
temperature eects on the length of the arms, and imposed upon that are the fringe
shifts corresponding to the eects of absolute motion, as shown in gure 4.
This temperature eect can be removed by subtracting from the data in each case
a best t to the data of a + bk, {k = 0, 1, 2, .., 8} for the rst 0
0
to 180
0
part of each
rotation data set. Then multiplying by 0.02 for the micrometer thread calibration gives
26
2 4 6 8 10 12 14 16
Marker
10
15
20
25
30
M
i
c
r
o
m
e
t
e
r
Figure 4: Plot of micrometer readings for July 11 12: 00 hr (7: 00 ST) showing the
absolute motion induced fringe shifts superimposed on the uniform temperature induced
fringe drift.
the fringe-shift data points in gure 6. This factor of 0.02 converts the micrometer
readings to fringe shifts expressed as fractions of a wavelength. Similarly a linear t has
been made to the data from the 180
0
to 360
0
part of each rotation data set. Separating
the full 360
0
rotation into two 180
0
parts reduces the eect of the temperature drift not
being perfectly linear in time.
In the quantum-foam physics there are four main velocities that contribute to the
total velocity:
v = v
cosmic
+v
tangent
v
in
v
E
. (70)
Here v
cosmic
is the velocity of the Solar system through space, while the other three
are local Solar system eects: (i) v
tangent
is the tangential orbital velocity of the Earth
about the Sun, (ii) v
in
is a quantum-gravity radial in-ow of the quantum foam past the
Earth towards the Sun, and (iii) the corresponding quantum-foam in-ow into the Earth
is v
E
and makes no contribution to a horizontally operated interferometer, assuming the
velocity superposition approximation, and also that the turbulence associated with that
ow is not signicant. The minus signs in (70) arise because, for example, the in-ow
towards the Sun requires the Earth to have an outward directed velocity against that
in-ow in order to maintain a xed distance from the Sun, as shown in gure 5. The
superposition in (70) is justied by the analysis in section 2.9. For circular orbits v
tangent
and v
in
are given by
v
tangent
=

GM
R
, (71)
27
v
in
=

2GM
R
, (72)
while the net speed v
N
of the Earth from the vector sum v
N
= v
tangent
v
in
is
v
N
=

3GM
R
, (73)
where M is the mass of the Sun, R is the distance of the Earth from the Sun, and G is
Newtons gravitational constant. G is essentially a measure of the rate at which matter
eectively dissipates the quantum-foam. The gravitational acceleration arises from
inhomogeneities in the ow. These expressions give v
tangent
= 30km/s, v
in
= 42.4km/s
and v
N
= 52km/s.

?
A
A
A
A
A
A
AK
v
in
Sun
v
tangent
v
N
Figure 5: Orbit of Earth about the Sun dening the plane of the ecliptic with tangential
orbital velocity v
tangent
and quantum-foam in-ow velocity v
in
. Then v
N
= v
tangent
v
in
is the velocity of the Earth relative to the quantum foam, after subtracting v
cosmic
.
Figure 6 shows all the data for the 1887 Michelson-Morley experiment for the fringe
shifts after removal of the temperature drift eect for each averaged 180 degree rotation.
The dotted curves come from the best t of
0.4
30
2
k
2
air
v
2
P
cos(2( )) to the data. The
coecient 0.4/30
2
arises as the apparatus would give a 0.4 fringe shift, as a fraction of a
wavelength, with k = 1 if v
P
= 30 km/s [7]. Shown in each gure is the resulting value
of v
P
. In some cases the data does not have the expected cos(2( )) form, and so the
corresponding values for v
P
are not meaningful. The remaining ts give v
P
= 331 30
km/s for the 7 : 00 hr (ST) data, and v
P
= 328 50 km/s for the 13 : 00 hr (ST)
data. For comparison the full curves show the predicted form for the Michelson-Morley
data, computed for the latitude of Cleveland, using the Miller direction (see later) for
v
cosmic
of Right Ascension = 4
hr
54

and Declination = 70
0
30

and incorporating
the tangential and in-ow velocity eects for July. The magnitude of the theoretical
curves are in general in good agreement with the magnitudes of the experimental data,
excluding those cases where the data does not have the sinusoidal form. However there
28
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
7 hr July 11
296 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
13 hr July 12
374 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
7 hr July 11
355 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
13 hr July 12
144 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
7 hr July 9
343 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
13 hr July 09
147 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
7 hr July 9
125 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
13 hr July 09
275 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
7 hr July 8
156 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
13 hr July 8
313 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
7 hr July 8
181 km per sec
2 3 4 5 6 7 8 9
-0.04
-0.02
0
0.02
0.04
13 hr July 8
349 km per sec
Figure 6: Shows all the Michelson-Morley 1887 data after removal of the temperature
induced fringe drifts. The data for each 360
0
full turn (the average of 6 individual turns)
is divided into the 1st and 2nd 180
0
parts and plotted one above the other. The dotted
curve shows a best t to the data, while the full curves show the expected forms using
the Miller value for v
cosmic
, the tangential velocity and the in-ow velocity eect.
29
are signicant uctuations in the azimuth angle. These uctuations are also present in
the Miller data, and together suggest that this is a real physical phenomenon, and not
solely due to diculties with the operation of the interferometers.
The Michelson-Morley interferometer data clearly shows the characteristic sinusoidal
form with period 180
0
together with a large speed. Ignoring the eect of the refractive
index, namely using the Newtonian value of k = 1, gives speeds reduced by the factor k
air
,
namely k
air
v
P
= 0.0241 330km/s = 7.9 km/s. Michelson and Morley reported speeds
in the range 5km/s - 7.5km/s. These slightly smaller speeds arise because they averaged
all the 7:00 hr (ST) data, and separately all the 13:00 hr (ST) data, whereas here some
of the lower quality data has not been used. Michelson was led to the false conclusion
that because this speed of some 8 km/s was considerably less than the orbital speed of
30 km/s the interferometer must have failed to have detected absolute motion, and that
the data was merely caused by experimental imperfections. This was the awed analysis
that led to the incorrect conclusion by Michelson and Morley that the experiment had
failed to detect absolute motion. The consequences for physics were extremely damaging,
and are only now being rectied after some 115 years.
3.3 The Miller Interferometer Experiment: 1925-1926
Dayton Miller developed and operated a Michelson interferometer for over twenty years
with an eective arm length of L = 32m achieved by multiple reections. The steel
arms weighed 1200 kilograms and oated in a tank of 275 kilograms of Mercury. The
main sequence of observations being on Mt.Wilson in the years 1925-1926, with the
results reported in 1933 by Miller [8]. Miller developed his huge interferometer over the
years, from 1902 to 1906 in collaboration with Morley, and later at Mt.Wilson where the
most extensive interferometer observations were carried out. Miller was meticulous in
perfecting the operation of the interferometer and performed many control experiments.
The biggest problem to be controlled was the eect of temperature changes on the lengths
of the arms. It was essential that the temperature eects were kept as small as possible
but, so long as each turn was performed suciently quickly, any temperature eect could
be assumed to have been linear with respect to the angle of rotation. Then a uniform
background fringe drift could be removed, as in the Michelson-Morley data analysis (see
gure 4).
In all some 200,000 readings were taken during some 12,000 turns of the interferome-
ter. Analysis of the data involved the extraction of the speed v
M
and the azimuth angle
by eectively tting the observed time dierences, obtained from the observed fringe
shifts, using (68), i.e. with k = 1. Miller was of course unaware of the full theory of
the interferometer and so he assumed the Newtonian theory, which neglected both the
Fitzgerald-Lorentz contraction and the refractive index eects of the air eects.
Millers results for April, August and September 1925 and February 1926 are shown in
gure 7. Here the speeds shown are the proper speeds v
P
after correcting for the refractive
index eect, that is, by dividing Millers v
M
data values by k
air
=

(n
2
1) = 0.0241, as
in (69). Then for example a speed of v
M
= 10km/s gives v
P
= v
M
/k
air
= 415km/s. The
Miller data was rediscovered in 2002 at Case Western Reserve University, and that data
30
has been used in preparing gure 7. However this refractive index correction procedure
was not available to Miller. He understood that the theory of the Michelson interferom-
eter was not complete, and so he introduced the phenomenological parameter k in (58).
We shall denote his values by k. Miller then proceeded on the assumption that v should
have only two components: (i) a cosmic velocity of the Solar system through space, and
(ii) the orbital velocity of the Earth about the Sun. Over a year this vector sum would
result in a changing v, as was in fact observed, see gure 7. Further, since the orbital
speed was known, Miller was able to extract from the data the magnitude and direction
of v as the orbital speed oered an absolute scale. For example the dip in the v
P
plots
for sidereal times 16
hr
is a clear indication of the direction of v, as the dip arises at
those sidereal times when the projection v
P
of v onto the plane of the interferometer is
at a minimum. During a 24hr period the value of v
P
varies due to the Earths rotation.
As well the v
P
plots vary throughout the year because the vectorial sum of the Earths
orbital velocity v
tangent
and the cosmic velocity v
cosmic
changes. There are two eects
here as the direction of v
tangent
is determined by both the yearly progression of the Earth
in its orbit about the Sun, and also because the plane of the ecliptic is inclined at 23.5
0
to the celestial plane. Figure 8 show the expected theoretical variation of both v
P
and
the azimuth during one sidereal day in the months of April, August, September and
February. These plots show the clear signature of absolute motion eects as seen in the
actual interferometer data of gure 7.
Note that the above corrected Miller projected absolute speed of approximately v
P
=
415km/s is completely consistent with the corrected projected absolute speed of some
330km/s from the Michelson-Morley experiment, though neither Michelson nor Miller
were able to apply this correction. The dierence in magnitude is completely explained
by Cleveland having a higher latitude than Mt.Wilson, and also by the only two sidereal
times of the Michelson-Morley observations. So from his 1925-1926 observations Miller
had completely conrmed the true validity of the Michelson-Morley observations and was
able to conclude, contrary to their published conclusions, that the 1887 experiment had in
fact detected absolute motion. But it was too late. By then the physicists had incorrectly
come to believe that absolute motion was inconsistent with various relativistic eects
that had by then been observed. This was because the special relativity formalism had
been derived from the assumption that absolute motion was without meaning and so
unobservable in principle. Of course the earlier interpretation of relativistic eects by
Lorentz had by then lost out to this misunderstanding.
3.4 In-ow from the Miller Data
As already noted Miller was led to the conclusion that for reasons unknown the existing
theory of the Michelson interferometer did not reveal true values of v
P
, and for this
reason he introduced the parameter k, with k indicating his numerical values. Miller had
reasoned that he could determine both v
cosmic
and k by observing the interferometer
determined v
P
and over a year because the known orbital velocity of the Earth about
the Sun would modulate both of these observables, and by a scaling argument he could
determine the absolute velocity of the Solar system. In this manner he nally determined
31
0 5 10 15 20
100
200
300
400
500
600
September
0 5 10 15 20
100
200
300
400
500
600
February
0 5 10 15 20
100
200
300
400
500
600
April
0 5 10 15 20
100
200
300
400
500
600
August
Figure 7: Millers results from the 1925-1926 observations of absolute motion showing
the projected speed v
P
in km/s plotted against sidereal time in hours, after applying the
refractive index correction (69). The results are for April, August and September 1925
and February 1926. In most cases the results arise from observations extending over
much of each month, i.e not from a single day in each month. Therefore the data points
are not strictly in chronological order. The lines joining the data points are merely to
make the data points clearer. The uctuations in both v
P
, and in (not shown) appear
to be a combination of apparatus eects and genuine physical phenomena caused by
turbulence in the gravitational in-ow of space towards the Sun. Each data point arises
from analysis of the average of twenty full rotations of the interferometer. The theoretical
curves, which include the cosmic velocity, the tangential orbital velocity and the in-ow
eect, are shown for v
c
= 420km/s and ( = 4
hr
, = 80
0
), and are assembled in one
plot in gure 8 to display more clearly the diurnal and seasonal variations.
32
0 5 10 15 20
Sidereal Time
100
200
300
400
500
600
s
p
e
e
d
k
m
p
e
r
s
e
c
(a)
0 5 10 15 20
Sidereal Time
150
160
170
180
190
200
210
220
A
z
i
m
u
t
h
(b)
Figure 8: Expected theoretical variation of (a) the projected velocity v
P
, and (b)
the azimuth during one sidereal day in the months of April, August, September and
February, labelled by increasing dash length. These forms arise from a best t to the
data in gure 7. The cosmic speed is 420km/s in the direction ( = 5
hr
, = 80
0
), and
the tangential and in-ows velocities are as in (70). These plots show the characteristics
of the signature expected in observations of absolute motion.
that |v
cosmic
| = 208 km/s in the direction ( = 4
hr
54
m
, = 70
0
33

). However now
that the theory of the Michelson interferometer has been revealed an anomaly becomes
apparent. Table 3 shows v = v
M
/k
air
for each of the four epochs, giving speeds consistent
with the revised Michelson-Morley data. However table 3 also shows that k and the
speeds v = v
M
/k determined by the scaling argument are considerably dierent. Here the
v
M
values arise after taking account of the projection eect. That k is considerably larger
than the value of k
air
indicates that another velocity component has been overlooked.
Miller of course only knew of the tangential orbital speed of the Earth, whereas the new
quantum foam physics predicts that as-well there is a quantum-gravity radial in-ow v
in
of the quantum foam. We can re-analyse Millers data to extract a rst approximation
to the speed of this in-ow component. Clearly it is v
N
=
_
v
2
in
+ v
2
tangent
that sets the
scale and not v
tangent
, and because k = v
M
/v
tangent
and k
air
= v
M
/v
N
are the scaling
relations, then
v
in
= v
tangent

_
v
2
N
v
2
tangent
1,
= v
tangent

_
k
2
k
2
air
1. (74)
Using the k values in table 3 and the value of k
air
we obtain the v
in
speeds shown
in table 3, which give an average speed of 54 km/s, compared to the Newtonian in-ow
speed of 42 km/s. Note that the in-ow interpretation of the anomaly predicts that
k = (v
N
/v
tangent
) k
air
=

3 k
air
= 0.042. Of course this simple re-scaling of the Miller
results is not completely valid because (i) the direction of v
N
is of course dierent to
that of v
tangent
, and also not necessarily orthogonal to v
tangent
because of turbulence,
and (ii) also because of turbulence we would expect some contribution from the in-ow
33
Epoch v
M
k v = v
M
/k
air
v = v
M
/k v =

3v v
in
February 8 9.3 km/s 0.048 385.9 km/s 193.8 km/s 335.7 km/s 51.7 km/s
April 1 10.1 0.051 419.1 198.0 342.9 56.0
August 1 11.2 0.053 464.7 211.3 366.0 58.8
September 15 9.6 0.046 398.3 208.7 361.5 48.8
Table 3. The k anomaly: k k
air
= 0.0241, as the gravitational in-ow eect. Here v
M
and k come from tting the interferometer data, while v and v are computed speeds
using the indicated scaling. The average of the in-ow speeds is v
in
= 54 5 km/s,
compared to the Newtonian in-ow speed of 42 km/s. From column 4 we obtain the
average v = 417 40km/s. A t to the Miller speed data gives the curves in gures 7
and 8.
eect of the Earth itself, which has a speed of 11 km/s, namely that it is not always
perpendicular to the Earths surface, and so would give a contribution to a horizontally
operated interferometer.
An analysis that properly searches for the in-ow velocity eect clearly requires a
complete re-analysis of the Miller data, and this is now possible as the original data
sheets have been found. In gures 7 and 8 the results of a preliminary re-t are shown,
but a more extensive one is in progress. It should be noted that the direction diametrically
opposite ( = 4
hr
54
m
, = 70
0
33

), namely ( = 17
hr
, = +68

) was at one stage

considered by Miller as being possible. This is because the Michelson interferometer,
being a 2nd-order device, has a directional ambiguity which can only be resolved by
using the seasonal motion of the Earth. However as Miller did not include the in-ow
velocity eect in his analysis it is possible that a re-analysis might give this northerly
direction as the direction of absolute motion of the Solar system.
Hence not only did Miller observe absolute motion, as he claimed, but the quality and
quantity of his data has also enabled the conrmation of the existence of the gravitational
in-ow eect. This is a manifestation of the new theory of gravity and one which relates
to quantum gravitational eects via the unication of matter and space developed in [1].
As well the persistent evidence that this in-ow is turbulent indicates that this theory of
gravity involves self-interaction of space itself.
3.5 The Illingworth Experiment: 1927
In 1927 Illingworth [9] performed a Michelson interferometer experiment in which the
light beams passed through the gas Helium. Helium was used in the interferometer to
in fact reduce the temperature eects via any refractive index eects, as a good vacuum
was dicult to achieve. For Helium at STP n = 1.000036 and so k
2
He
= 0.00007, which
results in an enormous reduction in sensitivity of the interferometer. Nevertheless this
experiment gives an excellent opportunity to check the n dependence in (69). Illingworth,
not surprisingly, reported no ether drift to an accuracy of about one kilometer per
second. M unera [27] re-analysed the Illingworth data to obtain a speed v
M
= 3.13
1.04km/s. The correction factor in (69), 1/
_
n
2
He
1 = 118, is large for Helium and
gives v = 368 123km/s. As shown in gure 9 the Illingworth observations now agree
34
1 2 3 4 5
Experiment Code
100
200
300
400
500
600
5x 5x
5x
5x
Figure 9: Speeds v in km/s determined from various Michelson interferometer exper-
iments (1)-(4) and CMB (5): (1) Michelson-Morley (noon observations) and (2) (18
h
observations) see section 3.2, (3) Illingworth [9], (4) Miller, Mt.Wilson [8], and nally in
(5) the speed from observations of the CMB spectrum dipole term [19]. The results (1)-
(3) are not corrected for the 30km/s of the orbital motion of the Earth about the Sun
or for the gravitational in-ow speed, though these correction were made for (4) with the
speeds from table 3. The horizontal line at v = 369km/s is to aid comparisons with the
CMB frame speed data. The Miller direction is dierent to the CMB direction. Due to
the angle between the velocity vector and the plane of interferometer the results (1)-(3)
are less than or equal to the true speed, while the result for (4) is the true speed as this
projection eect was included in the analysis. These results demonstrate the remarkable
consistency between the three interferometer experiments. The Miller speed agrees with
the speed from the DeWitte non-interferometer experiment, in section 3.7. The lower
data, magnied by a factor of 5, are the original speeds v
M
determined from fringe shifts
using (58) with k = 1. This gure updates the corresponding gure in reference [5].
35
with those of Michelson-Morley and Miller, though they would certainly be inconsistent
without the ndependent correction, as shown in the lower data points (shown at 5
scale).
So the use by Illingworth of Helium gas has turned out have oered a fortuitous
opportunity to conrm the validity of the refractive index eect, though because of the
insensitivity of this experiment the resulting error range is signicantly larger than those
of the other interferometer observations. So nally it is seen that the Illingworth experi-
ment also detected absolute motion with a speed consistent with all other observations.
3.6 The New Bedford Experiment: 1963
In 1964 from an absolute motion detector experiment at New Bedford, latitude 42
0
N,
Jaseja et al [10] reported yet another null result. In this experiment two He-Ne masers
were mounted with axes perpendicular on a rotating table, see gure 10. Rotation of
the table through 90
0
produced repeatable variations in the frequency dierence of about
275kHz, an eect attributed to magnetorestriction in the Invar spacers due to the Earths
magnetic eld. Observations over some six consecutive hours on January 20, 1963 from
6: 00 am to 12: 00 noon local time did produce a dip in the frequency dierence of
some 3kHz superimposed on the 275kHz eect, as shown in gure 11 in which the local
times have been converted to sidereal times. The most noticeable feature is that the dip
occurs at approximately 17 18:00
hr
sidereal time (or 9 10:00 hrs local time), which
agrees with the direction of absolute motion observed by Miller and also by DeWitte
(see section 3.7). It was most fortunate that this particular time period was chosen as
at other times the eect is much smaller, as shown for example for the February data in
gure 7. The local times were chosen by Jaseja et al such that if the only motion was
due to the Earths orbital speed the maximum frequency dierence, on rotation, should
have occurred at 12:00hr local time, and the minimum frequency dierence at 6:00 hr
local time, whereas in fact the minimum frequency dierence occurred at 9:00 hr local
time.
As for the Michelson-Morley experiment the analysis of the New Bedford experiment
was also bungled. Again this apparatus can only detect the eects of absolute motion if
the cancellation between the geometrical eects and Fitzgerald-Lorentz length contrac-
tion eects is incomplete as occurs only when the radiation travels in a gas, here the
He-Ne gas present in the maser.
This double maser apparatus is essentially equivalent to a Michelson interferometer.
Then the resonant frequency of each maser is proportional to the reciprocal of the
out-and-back travel time. For maser 1

1
= m
V
2
v
2
2LV

1
v
2
c
2
, (75)
for which a Fitzgerald-Lorentz contraction occurs, while for maser 2

2
= m

V
2
v
2
2L
. (76)
36
-

6
?
?

Maser 2
Maser 1
PM
-
-

C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
CW
C
C
C
C
CW

v
PM
(a) (b)
Figure 10: Schematic diagram for recording the variations in beat frequency between
two optical masers: (a) when at absolute rest, (b) when in absolute motion at velocity v.
PM is the photomultiplier detector. The apparatus was rotated back and forth through
90
0
.
Here m refers to the mode number of the masers. When the apparatus is rotated the
net observed frequency dierence is = 2(
2

1
), where the factor of 2 arises as the
roles of the two masers are reversed after a 90
0
rotation. Putting V = c/n we nd for
v << V and with
0
the at-rest resonant frequency, that
= (n
2
1)
0
v
2
c
2
+ O(
v
4
c
4
). (77)
Including Fresenl drag in the masers changes the sign in (77), and so has no eect on
the analysis here. If we use the Newtonian physics analysis, as in Jaseja et al [10], which
neglects both the Fitzgerald-Lorentz contraction and the refractive index eect, then
we obtain =
0
v
2
/c
2
, that is without the n
2
1 term, just as for the Newtonian
analysis of the Michelson interferometer itself. Of course the very small magnitude of
the absolute motion eect, which was approximately 1/1000 that expected assuming
only an orbital speed of v = 30 km/s in the Newtonian analysis, occurs simply because
the refractive index of the He-Ne gas is very close to one. It is possible to compare the
refractive index of the He-Ne gas mixture in the maser with the value extractable from
this data: n
2
= 1 + 30
2
/(1000 400
2
), or n = 1.0000028. Nevertheless given that it
is small the sidereal time of the obvious dip coincides almost exactly with that of the
other observations of absolute motion.
The New Bedford experiment was yet another missed opportunity to have revealed
the existence of absolute motion. Again the spurious argument was that because the
Newtonian physics analysis gave the wrong prediction then the vacuum special relativity
must be correct. But the analysis simply failed to take account of the Fitzgerald-Lorentz
contraction, which had been known since the end of the 19
th
century. As well the authors
failed to convert their local times to sidereal times and compare the time for the dip
with Millers time.
37
14 15 16 17 18 19 20
Sidereal Time
265
267.5
270
272.5
275
277.5
280
Figure 11: Frequency dierence in kHz between the two masers in the 1963 New Bed-
ford experiment after a 90
0
rotation. The 275kHz dierence is a systematic repeatable
apparatus eect, whereas the superimposed dip at 17 18: 00
hr
sidereal time of ap-
proximately 3kHz is a real time dependent frequency dierence. The full curve shows
the theoretical prediction for the time of the dip for this experiment using the Miller
direction for v ( = 4
hr
54
m
, = 70
0
33

) with |v| = 417km/s and including the Earths

orbital velocity and Sun gravitational in-ow velocity eects for January 20, 1963. The
absolute scale of this theoretical prediction was not possible to compute as the refractive
index of the He-Ne gas mixture was unknown.
3.7 The DeWitte Experiment: 1991
The Michelson-Morley, Illingworth, Miller and New Bedford experiments all used Michel-
son interferometers or its equivalent in gas mode, and all revealed absolute motion. The
Michelson interferometer is a 2nd-order device meaning that the time dierence between
the arms is proportional to (v/c)
2
. There is also a factor of n
2
1 and for gases like
air and particularly Helium or Helium-Neon mixes this results in very small time dif-
ferences and so these experiments were always very dicult. Of course without the gas
the Michelson interferometer is incapable of detecting absolute motion, and so there are
fundamental limitations to the use of this interferometer in the study of absolute motion
and related eects.
In a remarkable development in 1991 a research project within Belgacom, the Bel-
gium telecommunications company, stumbled across yet another detection of absolute
motion, and one which turned out to be 1st-order in v/c. The study was undertaken by
Roland DeWitte [13]. This organisation had two sets of atomic clocks in two buildings
in Brussels separated by 1.5 km and the research project was an investigation of the task
of synchronising these two clusters of atomic clocks. To that end 5MHz radiofrequency
signals were sent in both directions through two buried coaxial cables linking the two
clusters. The atomic clocks were caesium beam atomic clocks, and there were three in
each cluster. In that way the stability of the clocks could be monitored. One cluster was
38
0 10 20 30 40 50 60 70
Sidereal Time
-15
-10
-5
0
5
10
15
20
n
s
Figure 12: Variations in twice the one-way travel time, in ns, for an RF signal to travel
1.5 km through a coaxial cable between Rue du Marais and Rue de la Paille, Brussels. An
oset has been used such that the average is zero. The denition of the sign convention
for t used by DeWitte is unclear. The cable has a North-South orientation, and the
data is dierence of the travel times for NS and SN propagation. The sidereal time for
maximum eect of 17hr (or 5hr) (indicated by vertical lines) agrees with the direction
found by Miller and also by Jaseja et al, but because of the ambiguity in the denition
of t the opposite direction would also be consistent with this data. Plot shows data
over 3 sidereal days and is plotted against sidereal time. See gure 13b for theoretical
predictions for one sidereal day. The time of the year of the data is not identied. The
uctuations are evidence of turbulence associated with the gravitational in-ow towards
the Sun. Adapted from DeWitte [13].
39
0 5 10 15 20
Sidereal Time
5
10
15
20
25
n
s
(a)
0 5 10 15 20
Sidereal Time
-15
-10
-5
0
5
10
15
20
n
s
(b)
Figure 13: Theoretical predictions for the variations in travel time, in ns, for one
sidereal day, in the DeWitte Brussels coaxial cable experiment for v
cosmic
in the direction
(, ) = (17.5
h
, 65
0
) and with the Miller magnitude of 417 km/s, and including orbital
and in-ow eects (but without turbulence). Shown are the results for four days: for the
Vernal Equinox, March 21 (shortest dashes), and for 90, 180 and 270 days later (shown
with increasing dash length). Figure (a) Shows change in one-way travel time t
0
nv
P
/c
for signal travelling from N to S. Figure (b) shows t, as dened in (78), with an oset
such that the average is zero so as to enable comparison with the data in gure 12. t
is twice the one-way travel time. For the direction opposite to (, ) = (17.5
h
, 65
0
) the
same curves arise except that the identication of the months is dierent and the sign
of t also changes. The sign of t determines which of the two directions is the actual
direction of absolute motion. However the denition of the sign convention for t used
by DeWitte is unclear.
in a building on Rue du Marais and the second cluster was due south in a building on Rue
de la Paille. Digital phase comparators were used to measure changes in times between
clocks within the same cluster and also in the propagation times of the RF signals. Time
dierences between clocks within the same cluster showed a linear phase drift caused
by the clocks not having exactly the same frequency together with short term and long
term noise. However the long term drift was very linear and reproducible, and that drift
could be allowed for in analysing time dierences in the propagation times between the
clusters.
Changes in propagation times were observed and eventually observations over 178
days were recorded. A sample of the data, plotted against sidereal time for just three
days, is shown in gure 12. DeWitte recognised that the data was evidence of absolute
motion but he was unaware of the Miller experiment and did not realise that the Right
Ascension for maximum/minimum propagation time agreed almost exactly with Millers
direction (, ) = (17.5
h
, 65
0
). In fact DeWitte expected that the direction of absolute
40
motion should have been in the CMB direction, but that would have given the data
a totally dierent sidereal time signature, namely the times for maximum/minimum
would have been shifted by 6 hrs. The declination of the velocity observed in this
DeWitte experiment cannot be determined from the data as only three days of data are
available. However assuming exactly the same declination as Miller the speed observed
by DeWitte appears to be also in excellent agreement with the Miller speed, which in
turn is in agreement with that from the Michelson-Morley and Illingworth experiments,
as shown in gure 9.
Being 1st-order in v/c the Belgacom experiment is easily analysed to sucient accu-
racy by ignoring relativistic eects, which are 2nd-order in v/c. Let the projection of the
absolute velocity vector v onto the direction of the coaxial cable be v
P
as before. Then
the phase comparators reveal the dierence between the propagation times in NS and
SN directions. First consider the analysis with no Fresnel drag eect,
t =
L
c
n
v
P

L
c
n
+ v
P
,
= 2
L
c/n
n
v
P
c
+ O(
v
2
P
c
2
) 2t
0
n
v
P
c
. (78)
Here L = 1.5 km is the length of the coaxial cable, n = 1.5 is the refractive index of
the insulator within the coaxial cable, so that the speed of the RF signals is approximately
c/n = 200, 000km/s, and so t
0
= nL/c = 7.5 10
6
sec is the one-way RF travel time
when v
P
= 0. Then, for example, a value of v
P
= 400km/s would give t = 30ns.
Because Brussels has a latitude of 51
0
N then for the Miller direction the projection
eect is such that v
P
almost varies from zero to a maximum value of |v|. The DeWitte
data in gure 12 shows t plotted with a false zero, but shows a variation of some 28 ns.
So the DeWitte data is in excellent agreement with the Millers data. There is ambiguity
in reference [13] as to whether the time variations in gure 12 include the factor of 2
or not, as dened in (78). It is assumed here that a factor of 2 is included. The Miller
experiment has thus been conrmed by a non-interferometer experiment if we ignore a
Fresnel drag.
But if we include a Fresnel drag eect then the change in travel time t
F
becomes
t
F
=
L
c
n
+ bv
P
v
P

L
c
n
bv
P
+ v
P
,
= 2
L
c
v
P
c
+ O(
v
2
P
c
2
),
=
1
n
2
t, (79)
where b = 1 1/n
2
is the Fresnel drag coecient. Then t
F
is smaller than t by a
factor of n
2
= 1.5
2
= 2.25, and so a speed of v
P
= 2.25 400 = 900 km/s would be
required to produce a t
F
= 30 ns. This speed is inconsistent with the results from
gas-mode interferometer experiments. This raises the question as to whether the Fresnel
41
eect is present in transparent solids, and indeed whether it has ever been studied? As
well we are assuming the conventional eletromagnetic theory for the RF elds in the
coaxial cable. An experiment to investigate this is underway at Flinders university.
The actual days of the data in gure 12 are not revealed in reference [13] so a detailed
analysis of the DeWitte data is not possible. Nevertheless theoretical predictions for
various days in a year are shown in gure 13 using the Miller speed of v
cosmic
= 417
km/s (from table 3) and where the diurnal eects of the Earths orbital velocity and the
gravitational in-ow cause the range of variation of t and sidereal time of maximum
eect to vary throughout the year. The predictions give t = 30 4 ns over a year
compared to the DeWitte value of 28 ns in gure 12. If all of DeWittes 178 days of data
were available then a detailed analysis would be possible.
Reference [13] does however reveal the sidereal time of the cross-over time, that is
a zero time in gure 12, for all 178 days of data. This is plotted in gure 14 and
demonstrates that the time variations are correlated with sidereal time and not local
solar time. A least squares best t of a linear relation to that data gives that the cross-
over time is retarded, on average, by 3.92 minutes per solar day. This is to be compared
with the fact that a sidereal day is 3.93 minutes shorter than a solar day. So the eect
is certainly cosmological and not associated with any daily thermal eects, which in any
case would be very small as the cable is buried. Miller had also compared his data against
sidereal time and established the same property namely that, up to small seasonal eects
caused by the Earths orbital plane being inclined to the equatorial plane, features in the
data tracked sidereal time and not solar time; see reference [8] for a detailed analysis.
The DeWitte data is also capable of resolving the question of the absolute direction
of motion found by Miller. Is the direction (, ) = (17.5
h
, 65
0
) or the opposite direction?
Being a 2nd-order Michelson interferometer experiment Miller had to rely on the Earths
seasonal eects in order to resolve this ambiguity, but his analysis of course did not
take account of the gravitational in-ow eect, and so until a re-analysis of his data his
preferred choice of direction must remain to be conrmed. The DeWitte experiment
could easily resolve this ambiguity by simply noting the sign of t. Unfortunately it is
unclear in reference [13] as to how the sign in gure 12 is actually dened, and DeWitte
does not report a direction expecting, as he did, that the direction should have been the
same as the CMB direction ( = 11.20
h
, = 7.22
0
).
3.8 The Torr-Kolen Experiment: 1981
A coaxial cable experiment similar to but before the DeWitte experiment was performed
at the Utah University in 1981 by Torr and Kolen [12]. This involved two rubidium vapor
clocks placed approximately 500m apart with a 5 MHz sinewave RF signal propagating
between the clocks via a nitrogen lled coaxial cable maintained at a constant pressure of
2 psi. This means that the Fresnel drag eect is not important in this experiment. Un-
fortunately the cable was orientated in an East-West direction which is not a favourable
orientation for observing absolute motion in the Miller direction, unlike the Brussels
North-South cable orientation. There is no reference to Millers result in the Torr and
Kolen paper, otherwise they would presumably not have used this orientation. Never-
42
0 25 50 75 100 125 150 175
Local Time days
100
200
300
400
500
600
700
m
i
n
u
t
e
s
Figure 14: Plot of the negative of the drift of the cross-over time between minimum
and maximum travel-time variation each day (at 10
h
1
h
ST) versus local solar time
for some 180 days. The straight line plot is the least squares t to the experimental data,
giving an average slope of 3.92 minutes/day. The time dierence between a sidereal day
and a solar day is 3.93 minutes/day. This demonstrates that the eect is related to
sidereal time and not local solar time. The actual days of the year are not identied in
reference [13]. Adapted from DeWitte [13].
theless there is a projection of the absolute motion velocity onto the East-West cable
and Torr and Kolen did observe an eect in that, while the round speed time remained
constant within 0.0001%c, typical variations in the one-way travel time were observed,
as shown in gure 15 by the data points. The theoretical predictions for the Torr-Kolen
experiment for a cosmic speed of 417 km/s in the direction (, ) = (17.5
h
, 65
0
), and
including orbital and in-ow velocities, are shown in gure 15. As well the maximum
eect occurred, typically, at the predicted times. So the results of this experiment are
also in agreement with the Miller direction, and the speed of 417 km/s which of course
only arises after re-scaling the Miller speeds for the eects of the gravitational in-ow.
As well Torr and Kolen reported uctuations in both the magnitude and time of the
maximum variations in travel time just as DeWitte observed some 10 years later. Again
we argue that these uctuations are evidence of genuine turbulence in the in-ow as
discussed in section 3.10. So the Torr-Kolen experiment again shows strong evidence for
the new theory of gravity, and which is over and above its conrmation of the various
observations of absolute motion.
3.9 Galactic In-ow and the CMB Frame
Absolute motion (AM) of the Solar system has been observed in the direction ( =
17.5
h
, = 65
0
), up to an overall sign to be sorted out, with a speed of 417 40 km/s.
This is the velocity after removing the contribution of the Earths orbital speed and the
Sun in-ow eect. It is signicant that this velocity is dierent to that associated with the
43
0 5 10 15 20
Sidereal Time - hours
-3
-2
-1
0
1
2
3
4
P
h
a
s
e
D
i
f
f
e
r
e
n
c
e
-
n
s
Figure 15: Data from the 1981 Torr-Kolen experiment at Logan, Utah [12]. The data
shows variations in travel times (ns), for local times, of an RF signal travelling through
500m of coaxial cable orientated in an E-W direction. Actual days are not indicated
but the experiment was done during February-June 1981. Results are for a typical day.
For the 1st of February the local time of 12:00 corresponds to 13:00 sidereal time. The
predictions are for March (shortest dashes) and June, for a cosmic speed of 417 km/s
in the direction (, ) = (17.5
h
, 65
0
), and including orbital and in-ow velocities but
without theoretical turbulence.
44
Cosmic Microwave Background (CMB) relative to which the Solar system has a speed
of 369 km/s in the direction ( = 11.20
h
, = 7.22
0
), see [19]. This CMB velocity
is obtained by nding the preferred frame in which this thermalised 3
0
K radiation is
isotropic, that is by removing the dipole component. The CMB velocity is a measure
of the motion of the Solar system relative to the universe as a whole, or atleast a shell
of the universe some 14Gyrs away, and indeed the near uniformity of that radiation in
all directions demonstrates that we may meaningfully refer to the spatial structure of
the universe. The concept here is that at the time of decoupling of this radiation from
matter that matter was on the whole, apart from small observable uctuations, at rest
with respect to the quantum-foam system that is space. So the CMB velocity is the
motion of the Solar system with respect to space universally, but not necessarily with
respect to the local space. Contributions to this global CMB velocity arise from the
orbital motion of the Earth in the Solar system (this contribution is apparent in the
CMB observational data and is actually removed in the analysis), the orbital motion of
the Solar system within the Milky Way galaxy, giving a speed of some 230 km/s giving
together with local motion of the Solar system in the Milky Way, a net speed of some
250 km/s, and contributions from the motion of the Milky Way within the local cluster,
and so on to perhaps larger clusters.
On the other hand the AM velocity is a vector sum of this global velocity and the
net velocity associated with the local gravitational in-ows into the Milky Way and into
the local cluster. This is because the observation of the CMB velocity does not pick
up the local gravitational in-ows. Only gravitational lensing could aect that result,
and that is an extremely small eect within the Milky Way. If the CMB velocity had
been identical to the AM velocity then the in-ow interpretation of gravity would have
been proven wrong. We therefore have three pieces of experimental evidence for this
interpretation (i) the refractive index anomaly discussed previously in connection with
the Miller data, (ii) the turbulence seen in all detections of absolute motion, and now
(iii) that the AM velocity is dierent in both magnitude and direction from that of the
CMB velocity.
That the AM and CMB velocities are dierent contributes to the explanation oered
herein for the resolution of the dark matter problem. Rather than the galactic velocity
anomalies being caused by undiscovered dark matter we see that the in-ow into non
spherical galaxies, such as the spiral Milky Way, will be non-Newtonian. As well it
will be interesting to determine, at least theoretically, the scale of turbulence expected in
galactic systems, particularly as the magnitude of the turbulence seen in the AM velocity
is somewhat larger than might be expected from the Sun in-ow alone. Any theory for
the turbulence eect will certainly be checkable within the Solar system as the time scale
of this is suitable for detailed observation.
3.10 Gravitational Waves
The velocity ow-eld equation is expected to have solutions possessing turbulence, that
is, uctuations in both the magnitude and direction of the gravitational in-ow com-
ponent of the velocity ow-eld. Indeed all the gas-mode Michelson interferometer ex-
45
0 10 20 30 40 50 60 70
Sidereal Time
-75
-50
-25
0
25
50
75
100
T
u
r
b
u
l
e
n
c
e
k
m
p
e
r
s
Figure 16: Speed uctuations determined from gure 12 by subtracting a least squares
best t of the forms shown in gure 13b. A 1ns variation in travel time corresponds
approximately to a speed variation of 27km/s. The larger speed uctuations actually
arise from a uctuation in the cross-over time, that is, a uctuation in the direction of
the velocity. This plot implies that the velocity ow-eld is turbulent. The scale of this
turbulence is comparable to that evident in the Miller data, as shown in gure 7.
periments and coaxial cable experiments showed evidence of such turbulence. The rst
clear evidence was from the Miller experiment, as shown in gure 7. Miller oered no
explanation for these uctuations but in his analysis of that data he did running time
averages. Miller may have in fact have simply interpreted these uctuations as purely in-
strumental eects. While some of these uctuations may be partially caused by weather
related temperature and pressure variations, the bulk of the uctuations appear to be
larger than expected from that cause alone. Even the original Michelson-Morley data in
gure 6 shows variations in the velocity eld and supports this interpretation. However
it is signicant that the non-interferometer DeWitte data also shows evidence of turbu-
lence in both the magnitude and direction of the velocity ow eld, as shown in gure
16. Just as the DeWitte data agrees with the Miller data for speeds and directions the
magnitude uctuations, shown in gure 16, are very similar in absolute magnitude to,
for example, the speed turbulence shown in gure 7.
It therefore becomes clear that there is strong evidence for these uctuations being
evidence of physical turbulence in the ow eld. The magnitude of this turbulence ap-
pears to be somewhat larger than that which would be caused by the in-ow of quantum
foam towards the Sun, and indeed following on from section 3.9 some of this turbulence
may be associated with galactic in-ow into the Milky Way. This in-ow turbulence is
a form of gravitational wave and the ability of gas-mode Michelson interferometers to
detect absolute motion means that experimental evidence of such a wave phenomena has
been available for a considerable period of time.
46
4 Conclusions
Here extensive experimental evidence has been presented for the existence of a quantum-
foam substratum to space. Eects of motion through this substratum as well as ows
related to gravity are evident in this experimental data. The evidence suggests that
in fact the special relativity eects, which are well established by experiment, are be-
ing caused by absolute motion of systems through this quantum foam that is space.
This amounts to an experimental conrmation of the Lorentzian interpretation of such
relativistic eects. As well a new theory of gravity has been proposed based on a general-
isation of both the Newtonian and General Relativity theories of gravity. It passes all the
key existing tests, and as well also appears to be capable of explaining numerous grav-
itational anomalies. The phenomena present in these anomalies provide opportunities
for further tests of the new gravitational physics. The experimental data also supports
the conjecture herein that the gravitational ow displays turbulence. This amounts to
the discovery of a form of gravitational wave. The new theory of gravity suggests that
Newtonian gravity is only strictly applicable to cases of high spherical symmetry, such as
the case of the Solar system which is dominated by the massive central star. In the case
of a highly non-spherical spiral galaxy the new theory predicts gravitational forces very
dierent from those predicted by the Newtonian theory. The use of the Newtonian theory
in such cases has been used to argue for the existence of dark matter. The new theory
of gravity does away with the need for this concept of dark matter, which would also
be consistent with the fact that after extensive searches no such dark matter has been
detected. Because General Relativity was constructed to agree with Newtonian gravity
in the limit of low masses and low speeds, the apparent failure of the Newtonian theory
in these cases casts serious doubt on the validity of the General Relativity formalism
itself.
5 References
References
[1] Cahill R T 2002 Process Physics: Inertia, Gravity and the Quantum, Gen. Rel. and
Grav. 34, 1637-1656
[2] Cahill R T and Klinger C M 2000 Self-Referential Noise and the Synthesis of Three-
Dimensional Space, Gen. Rel. and Grav. 32(3), 529
[3] Cahill R T and Klinger C M 2000 Self-Referential Noise as a Fundamental Aspect
of Reality, Proc. 2nd Int. Conf. on Unsolved Problems of Noise and Fluctuations
(UPoN99), Eds. D. Abbott and L. Kish, Adelaide, Australia, 11-15th July 1999,
Vol. 511, p. 43 (American Institute of Physics, New York)
[4] Cahill R T, Klinger C M and Kitto K 2000 Process Physics: Modelling Reality as
Self-Organising Information, The Physicist 37(6), 191
47
[5] Cahill R T and Kitto K 2003 Michelson-Morley Experiments Revisited and the Cos-
mic Background Radiation Preferred Frame, Apeiron 10, No.2. 104-117
[6] Chown, M 2000 Random Reality, New Scientist, Feb 26, 165, No 2227, 24-2
[7] Michelson A A and Morley E W 1887 Philos. Mag. S.5, 24, No. 151, 449-463
[8] Miller D C 1933 Rev. Mod. Phys. 5, 203-242
[9] Illingworth K K 1927 Phys. Rev. 30, 692-696
[10] Jaseja T S, Javan A, Murray J and Townes C H 1964 Test of Special Relativity or
Isotropy of Space by Use of Infrared Masers, Phys. Rev. A 133, 1221
[11] Michelson, A.A. 1981 Amer. J. Sci. S. 3 22, 120-129
[12] Torr D G and Kolen P 1984 Precision Measurements and Fundamental Constants,
B.N. Taylor and W.D. Phillips, Eds. Natl. Bur. Stand.(U.S.), Spec. Publ. 617, 675
[13] DeWitte R http://www.ping.be/pin30390/
[14] Joos G 1930 Ann. d. Physik [5], 7, 385
[15] Kennedy H P and Thorndike E M 1932 Phys. Rev. 42 400-418
[16] Brillet A and Hall J L 1979 Phys. Rev. Lett. 42, No.9, 549-552
[17] Braxmaier C, M uller H, Pradl O, Mlynek J, Peters A and Schiller S 2002 Phys. Rev.
Lett. 88, 010401
[18] Lipa J A, Nissen J A, Wang S, Striker D A and Avalo D 2003 Phys. Rev. Lett. 90,
060403-1
[19] Lineweaver C et al. 1996 Astrophysics J. 470, 38
[20] Ives H 1939 J. Opt. Soc. Am. 29, 183(1939); 38, 413
[21] Kirkwood R L 1953 The Physical Basis of Gravitation, Phys. Rev. 92(6), 1557
[22] Kirkwood R L 1954 Gravitational Field Equations, Phys. Rev. 95(4), 1051
[23] Panleve P 1921 C. R. Acad. Sci., 173, 677
[24] Gullstrand A 1922 Ark. Mat. Astron. Fys., 16, 1
[25] Hicks W M 1902 Phil. Mag, [6], 3, 256, 555; 9, 555
[26] Schwarz J P, Robertson D S, Niebauer T M and Faller J E 1998 Science 282, 2230
[27] Munera H A 1998 Aperion 5, No.1-2, 37-54